Here's an anecdote to take to the bank. I have a Rev D ][+ board that was previously damaged and malfunctioned. A while back, I restored it, tested it, did a full burn in, RAM tests, the lot, and it passed them all. I put it into a static bag and put it on a shelf.
Today, when I was putting a board in a system, I selected that one first, as I had marked it 'Restored, 100%, 48K'. After I completely installed the board, I powered it up, and.... Screen of rubbish characters. Might be the 9334, but my point in all of this, is that I did not first bench test it again ere installing it. IDK what would have gone bad in two months sitting on a shelf, but I marked the bag 100% burn in verified--and this board had some suspect issues originally where it had phantom problems that some standard chip swaps seemed to cure.
The takeaway, is even if you tested something two months earlier, test it again ere spending the time installing it.